PERFUME is 5 year ERC Advanced funded project starting in Fall 2015. Electronic, communicating, devices have become critical to our daily lives and will increasingly impact the development of our economies and societies. Faced with the heterogeneity in the nature of such connected objects (from smartphones, tablets, wireless relays, sensors, to autonomous robot-like devices) and the growing demands for better quality-of-service and greater data traffic volumes, future communication networks will require more intelligent organization solutions that what is currently available for traditional cellular communications. On the bright side, our devices are becoming locally smarter and powerful (more computing power to run data intensive applications, more memory, more sensing capability). Unfortunately, these amazing capabilities are geared at local, individual performance, rather than maximizing performance collectively, as a network. In response, PERFUME envisions a radical shift in network design and aims at enabling collective intelligence among devices (such as for instance the growing number of objects which are connected to the internet) by tapping on mutual (although partial, limited) communication capabilities between them. PERFUME combines several fields of theoretical expertise:
  • Information and communication theory
  • Optimization, game, and coordination theory
  • Team decision theory
  • Machine Learning
  • Autonomous control theory
Such fields are enhanced with an experimentation effort based on EURECOM's Mobile Communications Dept.'s OpenAirInterface platform and dedicated to:
  • Autonomous flying relays
  • Coordination between wirelessly connected UAVs
An abstract of the project is shown below. Several openings at the PhD, postdoc, researcher and engineer level are available.


Advances in theory, integration techniques and standardization have led to huge progress in wireless technologies. Despite successes with past and current (5G) research, new paradigms leading to greater spectral efficiencies and intelligent network organizations will be in great demand to absorb the continuous growth in mobile data. With few exceptions such as ad-hoc topologies, classical wireless design places the radio device under the tight control of the network. Promising technologies envisioned in 5G such as (i) Coordinated MultiPoint (CoMP) techniques, (ii) Massive MIMO, or (ii) Millimeter-wave (mmWave) by-and-large abide by this model. Pure network-centric designs, such as optical cloud-supported ones raise cost and security concerns and do not fit all deployment scenarios. Also they make the network increasingly dependent on a large amount of signaling and device-made measurements. Many 6G scenarios will require new design approches where some intelligence and decision making is kept at the edge nodes and terminals, such as 6G-assisted autonomous car networks, drone-enabled networks, and other scenarios where nodes at the edge of the network feature robots or even simply AI-enhanced user terminals.

Team Decision for Communication Networks

PERFUME envisions a new approach to designing the communication networks, which taps into devices' new capabilities. Our approach recasts devices as distributed computational nodes solving together multi-agent problems, allowing to maximize the network performance by exploiting local measurement and information exchange capabilities. The success of the project relies on the understanding of new information theory limits for systems with decentralized information, the development of novel device communication methods, jointly with advanced team-based statistical signal processing algorithms. PERFUME will evidence the potential gains associated with exploiting the devices’ collective, network friendly intelligence, in pushing the frontiers of mobile Internet performance.

PERFUME is organized around a number of study items and subtasks:
  • Exploring the fundamental limits of device coordination
    • Limits of team decision in wirless networks
    • Optimal team communication design
  • Device coordination applied to the future Mobile Internet
    • Robust coordinated transmission methods
    • Coordinated massive MIMO systems
    • Device coordination based on millimeter wave communications
  • Real-life experimentations
    • Testbed developments (based on OpenAirInterface opensource protocol stack)
Prof. Gesbert's Webinar at the One World SP seminar series on Aug 24, 2020 "Learning to team play" focussed on the potential of machine learning methods to solve challenging decentralized coordination/team decision problems under arbitrary network information uncertainties. It can only be played directly on YouTube, click on the picture or follow the link below:



Autonomous Aerial Connected Devices

PERFUME pushing the development of algorithms and prototypes for autonomous aerial wireless devices, with a combination of machine-learning and signal processing fueled solutions together with custom hardware. PERFUME is envisioning the use of flying robots (autonomous drones) that are fully integrated with wireless network equipment in a number of scenarios. Flying relays or flying base stations can enable better and faster response in disaster recovery scenarios (after earthquakes or floods). They can also be used for intelligent IoT data collection, in monitoring smart cities and smart agriculture. Another use case is intelligent transport systems (ITS) where flying connect devices can be used in network-assisted autonomous driving scenarios.

Experimental and prototyping development

As part of this research program on autonomous decision-making wireless flying devices, PERFUME has been developing prototypes for "autonomous aerial cellular relay robots". These are custom-built aerial robots (a.k.a. UAV or micro-drones) which use machine learning algorithms designed in the context of PERFUME to self-optimize their parameters for improving overall radio network performance. An example lies in the optimization of their own position in 3D at all times based on selected radio measurements. At the radio-optimized positions, the autonomous AUV acts as a cellular relay which is capable of providing application-layer enhanced (LTE, 5G) connectivity to mobile users carrying off-the-shelf commercial terminals. The communication layers builds on algorithms leveraging the world's leading opensource implementation of cellular LTE (OpenAirInterface), developed at EURECOM's Communications Systems Department.

PERFUME team recently unveiled a first video demonstrating such a real-life device in an initial single-UAV single-user scenario: